Synthetic multi-view generation breaks the performance ceiling of single-view robotic datasets.
March 31, 2026
Original Paper
Beyond Viewpoint Generalization: What Multi-View Demonstrations Offer and How to Synthesize Them for Robot Manipulation?
arXiv · 2603.26757
The Takeaway
The paper demonstrates that single-view robotic policies reach a saturation point where more data doesn't help, but augmenting with synthesized novel views via the RoboNVS framework allows policies to continue scaling. This provides a blueprint for scaling robot manipulation using monocular real-world data.
From the abstract
Does multi-view demonstration truly improve robot manipulation, or merely enhance cross-view robustness? We present a systematic study quantifying the performance gains, scaling behavior, and underlying mechanisms of multi-view data for robot manipulation. Controlled experiments show that, under both fixed and randomized backgrounds, multi-view demonstrations consistently improve single-view policy success and generalization. Performance varies non-monotonically with view coverage, revealing eff